Another ‘Samuelson, 1948’ moment? Evidence from machine learning | Hacker News Reader

Another ‘Samuelson, 1948’ moment? Evidence from machine learning

voxeu.org

49 pointshhs22 comments

Gettin somewhat astray from the original intention of the arcticle:

Macroeconomics still proclaims the BIP as the ultima ratio to measure economic performance. The BIP has been invented in times of extremely dire social and economic conditions, where deficit spending has been considered as the last resort to prevent a nation from collapsing. It proved to be a meaningful number during and after WWII where living on credit was somewhat necessary.

BIP measures growth and favors deficit. But we know that unbounded growth is neither healthy nor possible. I think the biggest problem is that we still try to shoehorn every economy into this single number, while it would make more sense to switch to another measure, depending on the operational mode a country/economic region/social system is operating in.

Maybe change operational economic mode according to the https://en.wikipedia.org/wiki/Kondratiev_wave ?

BIP?

Probably 'Bruttoinlandsprodukt', German for GDP.

>BIP measures growth and favors deficit. But we know that unbounded growth is neither healthy nor possible.

GDP growth doesn't mean using more resources. If I find a way to cut hair twice as fast, that's GDP growth. If somebody makes a CPU that's faster and consumes less energy, that's GDP growth. If someone finds a way to distribute parcels more efficiently via a better heuristic for the travelling salesman problem, that's GDP growth.

GDP growth is only bounded if there's a hard limit on the amount of useful ideas people can have, and I think even if such a limit exists, we're extremely far away from reaching it.

GDP measures sale prices.

> If I find a way to cut hair twice as fast, that's GDP growth.

Only it makes people get more haircuts or their is a labor shortage you get a second job.

> If somebody makes a CPU that's faster and consumes less energy, that's GDP growth

Only if it inspires more CPU purchase or usage that costs more energy use overall.

>Only it makes people get more haircuts or their is a labor shortage you get a second job.

There are things that are productive that aren't jobs. You having more time to read HN probably has some productive value, because it teaches you things. It doesn't directly translate to an increase in GDP immediately, but it does do that over time. The hair dresser might end up learning new things during that free time, which would eventually lead to an increase in GDP.

>Only if it inspires more CPU purchase or usage that costs more energy use overall.

Or we get to use the same amount of energy to do more computation. There's a lot of things that can eat up more computation.

Yes, economists should put production front and center. What is being produced by economy? How effectively and how robustly? What is not just the production output, but also capacity? What are the risks to the supply chains? What commodities are more important than others from the risk perspective?

For example, see Steve Keen talking a bit about this: https://www.youtube.com/watch?v=HE-44ngyYoA

Not an economist, but I think the challenge with that is the assumption that we can rationalize those decisions ex-ante whereas the beauty of capitalism (!) is it's emergent ability to rationalize capital allocation and production organically, which we can only observe ex-post. Which is not to say we can't nudge it here and there...

I think the issue with the GDP is that it's kind of a jack of all trades, master of none. It has a lot of uses, none of which it excels at. Those who would replace it only consider one use of it. And furthermore they try often to shoehorn their politics into the replacement.

A few uses of the gdp: How well are the people doing? How much slack is there in the economy - if we wanted to focus on essentials how much luxury productions could be changed into necessities production? How much trade power do we have - if we want to import something, do we have anything desirable to export in return? Maybe there are more.

I prefer 4 good numbers over 1 bad number.

Oh certainly, I agree that is the best approach.

>But we know that unbounded growth is neither healthy nor possible.

Why is it unhealthy? From where I'm standing, this growth better be essentially unbounded for our purposes. This value's growth represents the advancement of our technology and society. If it stops growing then it indicates that we have stopped advancing.

I think this growth, for practical purposes, is unbounded. Even if we somehow hit a wall and are unable to advance our technology itself, we can start harnessing the resources in our solar system and engineer more complex systems to take care of our needs. The resources in our solar system might as well be infinite compared to what humanity uses so far.

To put it into perspective what a lack of growth means: 80-90% of all scientists are alive today. Before the industrial revolution, the estimated GDP per capita (PPP) doubled in 1500 years. That is to say, the average person in the area of Denmark in the year 100 AD could buy about 1-2 cups of 1990s coffee for their day's worth of labor, while the average person in the area of Denmark 1600 AD could buy about 2-4 cups of 1990s coffee for their day's worth of labor.[0] That is to say, life for the average person didn't improve much over the centuries. To compare this to the modern world (post industrial revolution): Denmark's GDP per capita PPP doubled from 1999 to 2018 - less than 20 years. The average Dane today could afford 80 cups of that 1990s coffee for a day's labor.

This "unbounded growth" is fundamental to our modern society.

[0] https://en.m.wikipedia.org/wiki/List_of_regions_by_past_GDP_...

> >But we know that unbounded growth is neither healthy nor possible.

> Why is it unhealthy?

Assuming unbounded growth with bounded resources?

There is always a struggle whether macroeconomic really exist or it is an aggregative only, except for Money Supply. You can have all your ISLM model but it assumed equilibrium for it to work. But real work never have that (and hence you have money which does not need in a global perfect equilibrium model).

What is this? Economics for ants?

These acquarium models of economy are so inadequate to describe real world it hurts. If the only thing we as economic agents supposedly do is borrowing and transacting, the central banks keep lowering interest rates to spur demand for goods. Unbeknownst to them there is this thing called real estate market and when they lower interest rates people start parking their hard earned money into it for safety instead of spending it for goods like crazy. It's only indirect effect of people feeling safer with their money parked in houses that they start spending more for goods.

But somehow land and real estate and how it affects economics is never included in these textbook economics for ants that are used globally at schools and central banks.

Land and real estate is not just some detail to be left out. It is actually the thing which defines employment and unemployment. In the old days it was customary to set up your own farm and self subsistence if you could get hold of enough land for it. You asked yourself: why would I have to get a job with a shitty boss yelling at me when I can grow my own corn?

It's only after all land had been taken that you no longer have that option. Back then unemployment was a choice - there was always a 'job' of cultivating your own land.

Let's see if the world wakes up from blindness of their economic religions and discover other explanations of people interactions.

Ideas around universal basic income seem to gain traction in times of crisis. Perhaps this is the time Henry George's land economics start shining again after 100 years.

The problem is that nowadays your land is not your land. You can't do whatever you want on it. Everything is heavily restricted by permits and regulations. Sure, you can farm, but you can't start a factory or mine the minerals on it. You can't build even a house on it without permission from someone else. All of this means that land is only a means of realistic production if you can navigate the arcane rules of the government. I agree that real estate is an important part of the economy, but for many it essentially is just an investment or rent generator (to lower the investment costs). Our modern economy is about making stuff and land use isn't directly related to it as much as it used to be.

While this may cramp your style, it also prevents your nextdoor neighbor from tearing down their house to start an open-pit mine... I don't think you'll encounter too many building code problems if you buy a parcel of remote land, if all you wanted was to live in a geodesic dome?

I don't disagree with the regulations. I'm saying that there are enough regulations that land doesn't directly translate to means of production anymore.

>I don't think you'll encounter too many building code problems if you buy a parcel of remote land, if all you wanted was to live in a geodesic dome?

But you probably aren't allowed to live there. You might be fine if you're entirely alone, but your kids would probably be taken away from you if you had kids.

I suspect you may be right but it would help to see a model demonstrating the impact of adding/removing land from the model

Georgists build upon on David Ricardo's Law of rent which provides a model of land supply and its effect on employment and wealth distribution.

https://youtu.be/yyv1xYDWAxk https://youtu.be/FOmImcdbjfw http://www.henrygeorge.org/pchp11.htm

The only thing more over-simplifying than aquarium models is people accusing economists of over-simplifying.

It is a vast field with thousands of subtopics.

If an economist showed up and mocked Torvalds for not using Apache to React.js, they'd look about as silly as 99% of people accusing economists of ignorance of the actual world.

Umm, I don’t think you actually read the article. Sam Bowles is the most accomplished radical economist in America. You’re acting like he’s a textbook neoclassical.

Reply on news.ycombinator.com